Feature
If artificial general intelligence (AGI) could arrive in 2025, enterprise teams have some work to do.
OpenAI Co-Founder and CEO Sam Altman casually predicted he’s “excited” for AGI in 2025 in a recent interview with Y Combinator. OpenAI defines AGI as “AI systems that are generally smarter than humans.”
Here, a group of AI leaders share how enterprises can get ready for AGI — and when they believe AGI may reach the market.
Looking to AGI
How Should Executives Prepare for AGI?
AI leaders have a “unique responsibility to steer the conversation about AGI toward positive, human-centric outcomes,” said Robin Patra, head of data, platform, product and engineering at ARCO Construction.
“Our focus must shift from whether AGI will arrive to how we prepare society, businesses and governments to coexist with it,” Patra said. “This involves continuous innovation, governance and above all, collaboration across sectors.”
Implement and Use AI Now
Enterprises should always be prepared for change, despite differing opinions on AGI’s timing, said Raj Krishnan, a director at Microsoft.
Before considering AGI, organizations should focus on using AI “as it currently exists,” developing solid plans to move from pilot projects to successful deployments, Krishnan said. With a foundation, the lessons learned will “help prepare organizations for a future in which AGI can be used to build better, more powerful AI systems.”
AI leadership can prepare for AGI by “bringing today’s plans for AI to the next level,” by looking at “how they can take advantage of the upside of AI while protecting against the risks,” said Nicolas Chapados, VP of research at ServiceNow.
Corporate investments, he said, in data quality, AI infrastructure and integrations will “continue to pay off in an AGI world.”
Identify AGI Use Cases and Prepare Data
To prepare for AGI, AI leaders need to "define the context and use cases for AGI,” said Krishna Subramanian, co-founder and COO of Komprise.
For instance, she said, enterprises could start with generative AI for chat, as it’s a” great humanistic interface,” but the knowledge it uses comes from predictive AI or heuristic-based solutions.
In terms of data, AI leaders need to define how corporate data can interface with AGI, preparing how data will be searched, curated, ingested and audited for AI data workflows, which “will become critical,” Subramanian said.
Establish AI Governance
AI leaders must proactively establish governance models that “align with AGI's potential capabilities,” including defining ethical boundaries, legal compliance and decision-making frameworks, Patra said.
For example, during the AI transformation Patra helped lead at BlackRock, the team integrated ethical decision-making into investment algorithms to “ensure fairness and regulatory compliance.” AGI, he said, will “demand even more robust systems to manage its autonomous decision-making capabilities.”
The implementation of robust AI governance frameworks, including substantial investment in safety and security, will be critical to “limit downside risk” and ensure that AI’s rapid pace of development “remains aligned with the organization’s and society’s goals and values,” Chapados said.
Build Up Cybersecurity
AI leaders preparing for much more advanced AI need to “think about defense,” said Jason Stanley, head of the trust and governance lab at ServiceNow.
As AI capabilities improve, “they arm malicious actors outside organizations with sophisticated tools for attacks,” Stanley said. Leaders in AI need to understand, he said, where their organization is vulnerable, how threat vectors and attack types are evolving and what tools and skills they need to defend themselves.
“Even in cases where AI leaders decide to take a conservative posture on AI adoption internally, they won’t have a choice but to rework their defensive tools and posture, which might imply needing to adopt new AI for security technologies,” Stanley said.
Develop Human-Centric AI
Leaders in AI should prioritize designing AGI systems that “align with human values and needs,” Patra said. Human-centric AI, he said, requires embedding explainability, transparency and bias mitigation at the core of AGI development. Leaders should establish feedback loops where AGI systems are regularly “audited and adjusted” based on real-world use cases.
Upskill the Workforce
Enterprises need to prepare their workforce to “collaborate with AGI,” including upskilling teams to use AGI as a tool “rather than fearing it as a competitor,” Patra said.
In the construction industry, for instance, Patra’s seen how “AI can augment human capabilities,” improving safety and efficiency. AGI, he said, will “amplify this dynamic but only if the workforce is ready to adapt.”
Leadership “at all levels” should assess the organizational impact of AGI and invest in workforce development, Chapados said.
Instead of focusing on when AGI may arrive, AI leaders should focus on when AI systems are “getting close to or better than human performance on the task areas key to their business,” Stanley said.
“This could be much sooner than when we achieve overall AGI,” Stanley said.
Related Article: Artificial General Intelligence: Jumping to the New Inference Market S-Curve
Will AGI Arrive in 2025?
Randall Hunt, CTO at Caylent, said “AGI isn't arriving in 2025.”
Krishnan with Microsoft said he’s “skeptical that AGI will arrive by 2025, unless we redefine and clarify what AGI truly entails and what significant changes we expect beyond current AI limitations.”
Patra at ARCO Construction said he’s skeptical, and “while the pace of AI advancements is staggering, AGI arriving by 2025 is highly ambitious.”
Today’s Models Aren’t There
Current AI models, while impressive, still rely on specialized architectures and training paradigms that “fall short of human-like reasoning,” Patra said. The leap from narrow AI to general AI, he said, involves scaling compute power and innovating in hardware and algorithms, “which may take longer than two years.”
Despite advancements in foundation AI models, current models struggle with “key hallmarks of AGI” — true contextual understanding, reasoning and long-term planning, Patra said.
Certain Applications May be Close
However, Patra said, “narrow AGI-like capabilities” in specific domains, such as autonomous supply chain optimization and predictive maintenance, may emerge in 2025.
“These systems will mimic aspects of AGI but lack the generality and autonomy associated with true AGI,” Patra said.
Even if the AI market approaches AGI capabilities, he said there will likely be significant regulatory and societal pushback delaying its deployment.
“Organizations and governments will need to address existential risks before widespread adoption,” Patra said.
AGI May Already be Here
The frontier of AI capabilities is advancing at variable speed, said Stanley with ServiceNow. AI systems, he said, are “already capable of human or super-human performance in some areas, while still lagging far behind in many other areas,” creating a “jagged frontier of capabilities.”
“For AI leaders inside organizations, we may already be close to human-like performance in some task areas they care deeply about,” Stanley said.
While AGI is an area of research, some subdomains, such as generative AI, have started to show significant AGI promise, “because they interact like humans,” said Subramanian with Komprise.
Related Article: This Professor Has a Path Toward Human-Level AI. It Isn’t Through Language
Will AGI Arrive After 2025?
Krishnan at Microsoft said “if AGI is defined as an AI system capable of performing any intellectual task a human can, essentially replacing humans, then AGI is still far off.”
Webinar
Sep
25
CMS Briefing: A Live Look at What’s Next in AI-Driven Platforms
Learn how leading organizations are using AI‑driven tools to publish faster, personalize smarter and stay secure.
Register
Webinar
On demand
Ready or Not: How Data-First Organizations Are Unlocking Agentforce Potential
Learn how to cut through the noise, activate Agentforce and build a Salesforce AI strategy that actually delivers.
Watch Now
Webinar
On demand
AI in Customer Service: Faster Resolutions, Happier Customers
Don’t let rising demand burn out your team. See how to build a smarter, more resilient support org.
Watch Now
Webinar
On demand
From Hype to High-Impact CX Strategies That Actually Scale
Turn buzzworthy AI and outsourcing trends into measurable CX wins with fresh 2025 data.
Watch Now
Webinar
On demand
Insights to Action Rethinking the Contact Center for Real Business Impact
Join our exclusive webinar to hear CX executives share their innovative strategies for transforming service delivery.
Watch Now
Webinar
On demand
Accelerating Healthcare Ops with AI
Sign up to see how your peers are putting AI to work — and making it stick.
Watch Now
Recent developments in AI research have led to “skepticism about bold predictions for AGI, which now seem overly optimistic, especially given the perceived slowdown in AI model advancements,” Krishnan said.
AGI will take “a few more years to develop,” because even in closer sub-segments, such as generative AI, organizations are “still experimenting with how and where they can get the most value,” said Subramanian with Komprise.
Is AGI 5 Years Away?
Patra at ARCO Construction said he foresees “AGI emerging closer to 2030 or beyond.”
The journey to AGI, he said, will likely “follow an incremental path,” with breakthroughs in transfer learning, causal inference and unsupervised learning.
“It’s unlikely to be a single eureka moment but a series of interconnected advances across disciplines,” Patra said.
AGI development requires synthesizing cross-discipline insights from neuroscience, cognitive science and artificial intelligence, and the integration is “still in its infancy,” Patra said. For example, he said, models that emulate the human brain’s abstraction and generalization processes, such as neuromorphic computing, are “in early research stages.”
Patra said based on his experience leading global AI initiatives, “large-scale adoption of such disruptive technologies takes at least a decade to fully materialize.”
Half of AI Researchers Predict AGI by 2047
There’s “considerable uncertainty” on the arrival of “human-level AI,” said Chapados with ServiceNow.
The “most credible assessment,” he said, on AGI’s timing is the report “Thousands of AI Authors on the Future of AI” by a team of researchers from UC Berkeley, the University of Bonn and University of Oxford. The 38-page report is based on a survey of over 2,700 AI researchers.
The report says 10% of AI researchers believe AI could outperform humans at all possible tasks by 2027, and 50% believe the moment could happen by 2047.
Related Article: Sam Altman: AI Will Replace 95% of Creative Marketing Work
